form-data-encoder

Patch Changes

• 247fb3e Thanks @octet-stream! - Remove removeComments from tsconfig.json

Major Changes

• 7f953c7 Thanks @octet-stream! - Bring back support for CJS via tsup. This package can be used with both ES and CJS modules

• 5a4cd63 Thanks @octet-stream! - Drop Node.js 16 support. Minimal required version now is Node.js 18

• 7f953c7 Thanks @octet-stream! - Make FileLike and FormDataLike types compatible with File and FormData

Update

• Fix FormDataLike types compatibility with undici (#19);
• Set compilerOptions.module value to node16 to meet TS requirements;

---

All changes: v3.0.0...v3.0.1

Remove

• Remove deprecated isFileLike() utility, use isFile() instead;
• Remove deprecated FormDataEncoder.getContentLength() method.

Update

• BREAKING: Drop Node 14.x support. Since 3.x the minimal required version is 16.5.x;
• Implement internal chunk utility to split value into chunks, if the value size is greater than 65536 bytes. This applied both to Files and string values, because FormDataEnocder supports AsyncIterable returned by File.stream();
• Expose package.json in esports field of package.json.

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v2.1.4...v3.0.0

Update

• Fix for browser compatibility: Add a fallback for ReadableStream that does not implement Symbol.asyncIterator method;
• Improve documentation for FormDataEncoder constructor.

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v2.1.3...v2.1.4

Update

• Replace LowercaseObjectKeys TS utility with two interfaces: RawHeaders and FormDataEncoderHeaders. This should resolve #8;

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v2.1.2...v2.1.3

Update

• Fix exports order (#11)

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v2.1.1...v2.1.2

Update

• Fix typings exports in package.json

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v2.1.0...v2.1.1

Update

• Fix a bug with NaN returned as Content-Length value when FormDataEncoder discovers entries without known length.

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v2.0.1...v2.1.0

Update

• Adjust minimal required Node.js version to 14.17

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v2.0.0...v2.0.1

Breaking

• Drop CommonJS support. The package is now ESM only;
• Remove Encoder from exports;
• Remove deprecated isFormDataLike helper. Use isFormData instead.

Update

• Minimal required Node.js version is 14.18;
• Headers in FormDataEncoder.headers property can be accessed in case-insensitive manner. TypeScript users still have correct TS typings, but only for original and lowercased keys (for example Content-Type and content-type);
• FormData instance will be preserved in array inside of FormDataEncoder instance (previously it held a reference to given FormData instance);
• Deprecate isFileLike helper. Use isFile instead;
• Deprecate FormDataEncoder.getContentLength() method. The content-length is now static value, so this method makes no sense. Use FormDataEncoder.contentLength or FormDataEncoder.headers["content-length"] instead.

Add

• Expose FormDataEncoderHeaders type.

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v1.7.2...v2.0.0

Update

• Revert headers bug fix from 1.8.0, because it introduced breaking change in FormDataEncoder.getContentLength signature and behaviour (See #15).

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v1.8.0...v1.9.0

Update

• Make isFileLike utility less strict when testing file-ish objects (See #14);
• Backport headers changes from v2.1.0
• Backport browsers compatibility fix from v2.1.4

Important notice: I no longer support 1.x branch! You should switch to 2.x if you can!

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v1.7.2...v1.8.0

Update

• Replace negative lookbehind with replacer function in normalizeValue utility, because it of the compatibility issues in Safari (See: https://github.com/octet-stream/form-data-encoder/issues/5)

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v1.7.1...v1.7.2

Update

• Return back isFormDataLike export and mark it deprecated.

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v1.7.0...v1.7.1

Update

• Use only lowercased alphabetical + numeric characters in createBoundary function to improve compatibility with existent Blob implementation as it's one of the supported targets;
• In-code documentation improvements;

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v1.6.0...v1.7.0

Add

• Add options argument for FormDataEncoder constructor;
• Add support for per part Content-Length header. It turned off by default, because it is not spec-compliant extension - the web clients does not include it. Use enableAdditionalHeaders option to enable this header:

import {Readable} from "stream"

import {FormDataEncoder} from "form-data-encoder"
import {FormData, File} from "formdata-node"

const form = new FormData()
const file = new File(["My hovercraft is full of eels"], "file.txt", {type: "text/plain"})

form.set("field", "Some field")
form.set("file", file)

// You can pass options as the second or third argument
const encoder = new FormDataEncoder(form, {enableAdditionalHeaders: true})

Readable.from(encoder).pipe(process.stdout)

Outputs:

--form-data-boundary-ZrZBNwvvIVAsVl6f
Content-Disposition: form-data; name="field"
Content-Length: 10

Some field
--form-data-boundary-ZrZBNwvvIVAsVl6f
Content-Disposition: form-data; name="file"; filename="file.txt"
Content-Type: text/plain
Content-Length: 29

My hovercraft if full of eels
--form-data-boundary-ZrZBNwvvIVAsVl6f--

To find more information about additional headers per form-data part, see:

• https://github.com/jimmywarting/FormData/issues/116
• https://github.com/whatwg/fetch/issues/1130

All changes: https://github.com/octet-stream/form-data-encoder/compare/v1.5.4...v1.6.0

Update

• Documentation improvements;
• Fixes for FileLike interface.

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v1.5.3...v1.5.4

Update

• Improvements for typings and documentation.

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v1.5.2...v1.5.3

Update

• Fix typings for FileLike type.

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v1.5.1...v1.5.2

Update

• Fix documentation to match API changes.

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v1.5.0...v1.5.1

Add

• Introduce FormDataEncoder.contentLength property.

Update

• Make sure readonly properties actually read-only in runtime.

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v1.4.4...v1.5.0

Update

• Internal improvements: Use fixed size for boundary because createBoundary function is private.

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v1.4.3...v1.4.4

Update

• Add normalization for non-File field's value following the spec.

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v1.4.2...v1.4.3

Update

• Use less ambiguous name for form data encoder class and exports: Encoder -> FormDataEncoder;
• Add separate tsconfig file for AVA;

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v1.4.1...v1.4.2

Update

• Fixes for readme.md;
• Cast Content-Length header to string.

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v1.4.0...v1.4.1

Update

• Slightly optimize createBoundary using bitwise A << 0 instead of Math.floor();
• Reduce package size by removing comments for JS output (but keep comments for .d.ts output).

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v1.3.0...v1.4.0

Update

• (internals) Simplify createBoundary string random generator.

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v1.2.0...v1.3.0

Add

• Introduce .values() and Symbol.iterator methods, so the Encoder can be converted to Blob:

import {Encoder} from "form-data-encoder"
import {Blob} from "fetch-blob"

const form = new FormData()

form.set("field", "FormData serialization using Blob + from-data-encoder.")

const encoder = new Encoder(form)

const options = {
  method: "post",
  body: new Blob(encoder, {type: encoder.contentType})
}

const response = await fetch("https://httpbin.org/post", options)

console.log(await response.json())

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v1.1.0...v1.2.0

Update

• Get rid of nanoid in favour of crypto.randomBytes and focus on Node.js support only, because browsers can handle encoding themselves.

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v1.0.1...v1.1.0

Update

• Add escaping for CR, LF and " symbols in name and filename params;
• Add an alternative example for Blob as conversion target;

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v1.0.0...v1.0.1

Add

• Expose isFileLike and isFormDataLike

Update

• Documentation improvements
• Loosen tests in isFormDataLike
• Make private members actually private

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v0.5.1...v1.0.0

Update

• Fix for 3rd example in documentation: node-fetch currently does not support async iterables, so the example is not valid when the encoder targeting fetch-blob@3

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v0.5.0...v0.5.1

Update

• Lowercase default boundary string lead characters so it can be used when targeting the Encoder output to Blob:

import {FormData} from "formdata-polyfill/esm-min.js"
import {blobFrom} from "fetch-blob/from.js"
import {Encoder} from "form-data-encoder"

import Blob from "fetch-blob"
import fetch from "node-fetch"

async function toBlob(form) {
  const encoder = new Encoder(form) // No need to use custom boundary with lowercased alphabetical characters anymore
  const chunks = []

  for await (const chunk of encoder) {
    chunks.push(chunk)
  }

  return new Blob(chunks, {type: encoder.contentType})
}

const fd = new FormData()

fd.set("name", "John Doe")
fd.set("avatar", await blobFrom("path/to/an/avatar.png"), "avatar.png")

const options = {
  method: "post",
  body: await toBlob(fd)
}

await fetch("https://httpbin.org/post", options)

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v0.4.0...v0.5.0

Update

• Put TextEncoder usage to the instance-wide scope and move CRLF and DASHES to Encoder constructor
• Behaviour compatibility fix: Do not try to guess Content-Type header for a File from its name property

Remove

• Because of behaviour changes in the Encoder, the mime-types package was removed from dependencies

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v0.3.0...v0.4.0

Update

• Improve Encoder#headers API (make it an actual property)
• Improve documentation and add more usage examples

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v0.2.0...v0.3.0

Update

• Documentation improvements
• Replace Buffer with Uint8Array + TextEncoder
• The Encoder#encode() method now yields Uint8Array instead of Buffer

---

All changes: https://github.com/octet-stream/form-data-encoder/compare/v0.1.0...v0.2.0

Initial release

See readme.md for documentation